State of the Innovation Ecosystem

The global startup economy now counts 5.4 million active companies with a combined unicorn valuation of $5 trillion, per Startup Genome’s Global Startup Ecosystem Report 2025 — though that combined value is down 31% from the prior report. Growth and survival aren’t the same thing, and the gap between them is where the real ecosystem story sits.

Survival by region

US Bureau of Labor Statistics data puts the five-year survival rate for new private-sector establishments at roughly 48-51%, regardless of sector — close to a coin flip. Europe has no equivalent single dataset, but its capital dynamics point to a different failure mode: more startups reaching year five undercapitalized rather than failing outright. Atomico’s research puts the region’s cumulative underinvestment at $375 billion over the past decade, and VC concentration in the US remains roughly double Europe’s. Undercapitalized survival looks different from failure in the data, but it produces a similar outcome — companies that plateau instead of scale.

Survival by sector

Sector matters more than geography. BLS data shows agriculture, forestry, fishing, and hunting startups have the strongest five-year survival, with just 29.4% failing by year five (6.9% in year one). Retail trade performs similarly well. At the other end, mining, quarrying, and oil and gas extraction see 59.8% failure by year five, and information-sector startups — software, media, publishing — aren’t far behind at 53.2%. Capital intensity and exposure to commodity or attention cycles both correlate with lower survival, independent of region.

Where the capital is actually going

Sector survival data lines up with capital flows. AI attracted 40% of all global VC dollars in the latest reporting period, up from 26% in 2021, and deep tech took 36% of European VC dollars — nearly double its share four years ago. Capital is also concentrating in fewer, larger bets: over the trailing four quarters, 77% of global VC capital landed in scale-up rounds of $100 million or more, while only 6% reached companies raising under $15 million.

Top ecosystems, mostly unchanged

Silicon Valley, New York, and London have held the top three spots in Startup Genome’s ranking since 2020, though London slipped from a tie with New York into third place this year. Silicon Valley’s ecosystem value has passed $3 trillion — nearly three times the next-largest ecosystem — a gap that reflects the depth of follow-on capital available once a company survives its first few years, not just the density of early ideas.

The unicorn gap as a survival proxy

Unicorn creation is a rough proxy for how many companies survive long enough to reach outlier scale, and the regional gap here is stark. The US produced 611 unicorns in 2025 against Europe’s 134, and of the 83 new unicorns minted globally so far this year, 52 came from the US versus just 14 across all of Europe. Concentration matters as much as the raw count: San Francisco and New York alone account for over 64% of the US total, while Europe’s leaders — the UK, France, and Germany — together make up just over 60% of the region’s total, spread across three separate legal and capital systems rather than one.

That distinction reinforces the sector data above: it isn’t that European companies fail faster. It’s that fewer of them reach the scale-up capital that turns survival into an outlier outcome.

What the data means for builders

Survival correlates more tightly with sector capital intensity and access to scale-up funding than with founder quality or market timing. That’s the benchmark worth tracking: not whether an ecosystem produces good ideas, but whether it has the follow-on capital to let those ideas survive past year five.

Three figures matter more than any ranking on their own: the roughly 50% five-year survival baseline that holds regardless of sector, the 20-30 point spread between the strongest and weakest sectors within that baseline, and the widening concentration of capital into scale-up rounds of $100 million or more. Together they say survival is no longer mainly a function of building something people want — it’s a function of reaching the capital pool that lets a company keep operating long enough to find out. DB tracks these figures on an ongoing basis because they change the calculus on where to build, when to raise, and how much runway a given sector realistically requires.

Sources